Multitask Fine-Tuning for Passage Re-Ranking Using BM25 and Pseudo Relevance Feedback.
Meoungjun KimYoungjoong KoPublished in: IEEE Access (2022)
Keyphrases
- pseudo relevance feedback
- document retrieval
- fine tuning
- multi task
- ranking functions
- learning to rank
- retrieved documents
- language model
- query specific
- retrieval model
- passage retrieval
- relevance model
- document collections
- information retrieval
- ranked list
- ad hoc retrieval
- average precision
- relevance feedback
- learning tasks
- language modelling
- retrieval effectiveness
- multi class
- feature selection
- evaluation metrics
- relevant documents
- transfer learning
- query expansion
- web search
- learning problems
- query terms
- term frequency
- document length
- evaluation measures
- vector space model
- language modeling
- ranking algorithm
- information retrieval systems
- feature vectors
- image retrieval
- pairwise
- reinforcement learning